Ann: Neuren receives response to request for Breakthrough Therapy, page-44

I can see why the average investor would be confused about this – and hence the polarised views about the FDA.

The announcement from the trial clearly stated:

"The clinical benefit in the trial encompassed core symptoms of Rett syndrome and was observed in both clinician and caregiver assessments"

So if clinical benefit was observed why no approval?

Take clinicians observing “clinical benefit”

From the figure relating to clinician assessment on average children in cohort 2 at 70mg/kg moved slightly to the “minimally worse” category (at 4.2) compared to baseline (no difference).

The placebo group also moved toward the “minimally worse” category at 4.4.

Where the scale is rated: 1= very much improved, 2 = much improved, 3 = minimally improved, 4 = no change, 5 = minimally worse, 6 = much worse, 7 = very much worse

So a 0.2 of a difference on a 7 point scale.

No clinician in the world would agree that clinical benefit was achieved here.

It is a very subtle play on the word “clinical”.

Now throw in some change figures and complicated statistics (permutation testing) add a bit Bayasian confusion and no-one knows whether they are coming or going.

Except for the FDA that is.

While a certain degree of liberalism in the investor level description of results is fine at the P2 exploratory stage when pushed too far it runs into the brick wall of” clinical benefit” actually having to mean something resembling a minimally clinically important difference (MCID). Where that is exactly who knows but a long way from 0.2.

Cortinaboy a few posts back.

“For many parameters the drug didn't even beat the placebo. Yes, for some others it did, including the primary endpoints, but numbers were low, variability was high and for some measures it failed to make a statistically significant difference.”

True Cortinaboy ... but the trick (if you can call it that) was the absence of clear data showing baseline status of the two groups. Failing to account for the correlation between intercept (baseline) and slope confounds the change statistics. That adjustment was done via the permutation testing rather than using covariates .... is my guess. Put more simply the clinician rating of change is probably the cleanest change measure here.